Error Modeling Of MEMS Gyroscopes And Its Application On Flight Vehicle Integrated Navigation

Comparing with conventional electro-mechanical gyroscope and electro-optical gyroscope, MEMS gyroscope has the advantages of low cost, small size, low weight and high reliability. In the past decades MEMS gyroscope has become the major subject of widely research and development and will continue to boost the evolution of navigation technologies. At present, basing on the process level, MEMS gyroscope is restricted in the field of low level navigation. One of the methods for increasing its available accuracy is error analysis, modeling and compensation.Error analysis, modeling and compensation of MEMS gyroscope and its application in navigation of flight vehicle are researched in this paper. Firstly, performance indices of MEMS gyroscope are summarized and its ideal dynamic model is analyzed. Secondly, the operation principle, performance and error characteristics of three main parts, vibratory structure, transducer and electrocircuit, of MEMS gyroscope are discussed thoroughly. Thirdly, basing on the above conclusions, error model of MEMS gyroscope is built and its compensation theory is discussed. Finally, basing on the above error model of MEMS gyroscope, a distributed integrated navigation simulation system and a prototype of low cost integrated navigation system for flight vehicle are designed, simulation and car test are carried out and the conclusions are given.The detailed contents in this paper are as followed:1) Performance indices of MEMS gyroscope are summarized. Ideal MEMS gyroscope can be taken as an equivalent two freedom mass-spring-damping system and its dynamic model is analyzed under two operation modes of open-loop and closed-loop. The typical bounds of characteristic parameters for describing MEMS gyroscope's structure are determined.2) Electrostatic actuator and capacitive detect technologies are studied. The operation mechanism of the common capacitor structure, plate, transverse comb and lateral comb, are analyzed and compared. Finally, qualitative analysis is given according to the cross error and bias which exist in the electrostatic actuator and capacitive sensor.3) Errors that existed in driven circuit are studied, it is indicated that the dominating error factor is cross error caused by the delay of phase. Operation principle and noise characteristics of thee fundamental detect circuit elements, unit gain buffer, tran-resistant amplifier and charge integrated amplifier, are analyzed and compared. It is indicated that the performance of charge integrated amplifier is the best. Operation principle and performance of four typical detect circuit, synchronous detect, switched capacitive detect, correlated double sampling areanalyzed and compared. It is indicated that the synchronous detect and correlated double sampling can eliminate the most parts of error sources, the dominating error sources of them is thermal noise of first stage of amplifier. The thermal noise is often several orders higher than the mechanical thermal noise.4) Two typical types of MEMS gyroscope, linear vibratory structure gyroscope and rotational vibratory gyroscope, are studied. The integrated dynamic models of them are deduced by the correlated dynamic characteristic. Structure imperfects, offset of center of mass and deflection of vibratory mass, are added into the above models by the form of non-diagonal items in asymmetry stiffness matrix and asymmetry damping matrix. Dynamic characteristics and operation mode control is analyzed basing on the simplified models with the methods of trajectory plot and parameter averaging. Finally, the primary and secondary error factors correlated with structure is classified and analyzed.5) Determinable error model, stochastic error model and modeling method are analyzed in detailed. Especially, stochastic error modeling methods are discussed, including time serial analysis and Allan variance method. The common error model equation is deduced by analysis of its primary and secondary error factors. Finally, to a specific type of MEMS gyroscope, error modeling and analysis are provided by the method of time serial analysis and Allan variance analysis respectively according to the static test data.6) Error compensation theories used for gyroscope are analyzed and compared. Especially, particle filtering theory which is the research focus at the present time is discussed in detail. The correlative filtering algorithm and the mixture filtering algorithm which can reduce computing consumption are deduced. Finally, a new cascade filtering structure is presented which not only can reduce computing consumption but also increase filtering robustness and the correlative filtering algorithm is also deduced.7) Basing on the above error model of MEMS gyroscope, a distributed integrated navigation simulation system and a prototype of low cost integrated navigation system for flight vehicle are designed, simulation and car test are carried out and the conclusions are given.